---
dataset_info:
  features:
  - name: id
    dtype: int64
  - name: language
    dtype: string
  - name: verdict
    dtype: string
  - name: problem_id
    dtype: string
  splits:
  - name: train
    num_bytes: 705630705
    num_examples: 13065846
  download_size: 101643454
  dataset_size: 705630705
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
task_categories:
- text-generation
tags:
- code-generation
- benchmark
- c++
- python
---

Dataset for [Can LLMs Generate High-Quality Test Cases for Algorithm Problems? TestCase-Eval: A Systematic Evaluation of Fault Coverage and Exposure](https://huggingface.co/papers/2506.12278)

Github: https://github.com/FlowRays/TestCase-Eval